JPH07152756A - Document processor - Google Patents

Abstract

PURPOSE:To easily execute maintenance and to reduce capacity by collecting dictionaries for respective purposes to one dictionary in processing such as KANA (Japanese syllabary)/KANJI (Chinese character) conversion, voice synthesis and voice recognition requiring dictionaries. CONSTITUTION:In the case of executing KANA/KANJI conversion, reading is inputted, notation corresponding to the reading is retrieved from a dictionary reading part 33 and a KANA sentence is converted into a KANA/KANJI mixed sentence. At the time of voice synthesis, a notation string is inputted, voice data corresponding to the input are retrieved from a notation part 34 and voice data are read out and outputted as voice. In the case of executing voice recognition, voice data are inputted and notation corresponding to the input is retrieved from a voice part 35 to convert the voice data into the notation. In the case of executing dictionary maintenance, any one of reading, notation and voice is inputted and a record corresponding to the input is adopted as an object to be maintained. Thereby the consistency of contents in the dictionary can be maintained and labor and time for the maintenance of the dictionary is made unnecessary.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document processing device using a dictionary, such as kana-kanji conversion.

[0002]

2. Description of the Related Art Conventionally, processing performed using a dictionary such as kana-kanji conversion processing has been widely performed. In such a process, for example, a kana-kanji conversion requires a dictionary for that purpose, and if voice synthesis or voice recognition is to be performed, a separate dictionary suitable for the process, such as a voice synthesis or voice recognition dictionary, is used. I needed a dictionary.

[0003]

As described above, the conventional processing described above requires a dictionary corresponding to the processing.
There is a problem that the storage capacity required to store the dictionary becomes large. Further, for example, in the case of dictionaries used for processing the same language such as kana-kanji conversion, voice synthesis, and voice recognition, the same word is often registered in each dictionary, and one registered content is corrected for one dictionary. Then, all the dictionaries had to be modified, and maintenance of the dictionaries was not easy.

The present invention has been made in view of the above-mentioned conventional example, and an object of the present invention is to provide a document processing apparatus in which the capacity of a dictionary used for a plurality of purposes is kept small and the maintenance thereof is facilitated.

[0005]

[Means for Solving the Problems] and

In order to achieve the above object, the document processing apparatus of the present invention has the following configuration.

That is, a plurality of attributes related to a word are registered as headings of the corresponding word, and a dictionary in which a plurality of attributes related to the same word are associated with each other And a means for performing the next process based on the attribute retrieved by the retrieval means.

With this configuration, it is possible to specify any one of the attributes from a dictionary having a plurality of attributes, subtract other attributes of the corresponding word, and use that to perform the next process. Therefore, only one dictionary can be used to reduce the size of the dictionary.

More preferably, it further comprises a designating unit for designating maintenance for the dictionary, and a maintenance unit for maintaining the attribute of the word retrieved by the retrieval unit as designated by the designating unit.

More preferably, the designating means designates addition or deletion of a word or change of an attribute and designation of an attribute of a target word, and the maintenance means in the case of addition. The specified attribute is registered as a headline in the dictionary, in the case of deletion, all attributes related to the word corresponding to the specified attribute are deleted from the dictionary, and in the case of change, the specified attribute is set to the specified value. change.

With this configuration, the attributes corresponding to one word can be collectively targeted for maintenance, the maintenance work can be saved, and the consistency of the dictionary can be maintained.

More preferably, the plurality of attributes include reading, notation and voice data, and the processing in the next stage includes kana-kanji conversion, voice synthesis and voice recognition.

With this configuration, kana-kanji conversion, voice synthesis and voice recognition can be performed using a single dictionary,
This dictionary is easy to maintain and requires a small capacity.

More preferably, the plurality of attributes include reading, notation and character recognition data, and the processing in the next stage includes character recognition and kana-kanji conversion.

With this configuration, character recognition and kana-kanji conversion can be performed using a single dictionary, which is easy to maintain and requires a small capacity.

More preferably, the plurality of attributes include reading and notation, voice data, and character recognition data, and the subsequent processing includes kana-kanji conversion, voice synthesis, voice recognition, and character recognition.

With this configuration, kana-kanji conversion, voice synthesis, voice recognition and character recognition can be performed using a single dictionary, and this dictionary is easy to maintain and requires a small capacity.

[0017]

[First Embodiment] An information processing system having a kana-kanji conversion / voice synthesis / voice recognition function, which is a first embodiment of the present invention, will be described below.

FIG. 2 is a block diagram showing the configuration of the information processing system of this embodiment.

Input data input from an input device 21, such as a keyboard for code input, a mouse for coordinate input, an image scanner for image input, and a microphone for voice input, is processed by a central processing unit (CPU) 22.
Is processed by C, and the processing result is stored in a storage device 24 such as a RAM or a disk memory and outputs an image
It is output from the output device 23 such as an RT, a printer, or a speaker that outputs sound. The input device 21 has a mechanism for digitizing a voice input. The input device 21 may be an input device such as an external storage device such as a disk device or a communication network.

FIG. 1 is a block diagram of modules included in the system of this embodiment. As shown in FIG. 1, one dictionary 4 is used by the kana-kanji conversion unit 1, the voice synthesis unit 2, and the voice recognition unit 3. The processing by each module is realized by the CPU 22 executing a program stored in the storage device 24. The dictionary 4 is stored in the storage device 24.

<Structure of Dictionary> FIG. 3 is a diagram showing the overall structure of an example of the dictionary 4. The dictionary 4 includes a character index 31, a voice index 32, a reading unit 33, a writing unit 34, and a voice unit 35. In the figure, a character index 31 is an index for retrieving a corresponding record from a key reading or notation, and a voice index 32 is an index for retrieving a corresponding record from key voice data. is there.

The reading part 33 is a part in which data for subtracting the notation from the kana or romaji reading is accumulated. As shown in FIG. 4A, the record (reading part word) has a reading string as a heading, a pointer connected to the record (notation part word) indicating the corresponding notation, a part of speech, and a conjugation form. It stores grammatical information such as.

The notation section 34 is a section in which data for subtracting voice data corresponding to the notation such as Chinese characters is accumulated. The record (notation part word) is shown in FIG.
As shown in (b), a notation string as a heading,
A pointer to a record (speech word) in which the corresponding voice data is stored, grammatical information of the record, and the like are stored.

The voice part 35 is a part in which data for subtracting a notation corresponding to the voice data from the voice data is accumulated. The record (speech word) is shown in FIG. 4 (c).
As shown in, the record that stores the voice data that serves as a headline and the corresponding notation (notation part word)
A pointer to and the grammatical information of the record are stored. The audio data may be a digitized audio signal.

As described above, the dictionary 4 is a combination of a plurality of dictionaries having the same registered items and different attributes regarding the corresponding items. That is, in the kana-kanji conversion, a dictionary is used for associating the reading and the notation, and in speech synthesis, the dictionary is used for associating the notation and the voice data. There is reading, notation, and voice 1
If there are three words, they will always exist. Therefore, if it is a dictionary that has words as items, the attributes, that is, reading, notation,
You can combine voices into one dictionary. Although the word is taken up as an item in the embodiment, the item is not limited to this.

<Kana-Kanji conversion> In Kana-Kanji conversion, the notation data corresponding to the reading data is searched from the reading data,
Kana-Kanji conversion is performed. FIG. 5 shows a flowchart of the processing procedure. This flowchart is based on the storage device 2 of FIG.
It is realized by the CPU 22 executing the program stored in 4.

First, in step S51, a reading string is input from the input device 21. The reading string is input in Roman characters or kana and is processed as a kana character string representing reading. Next, in step S52, the reading string is analyzed grammatically such as cutting out clauses and words. In step S53, the reading unit 33 of the dictionary 4 is searched for the word thus obtained using the reading as a key. Regarding the processing of homophones, there are known methods of processing such as determining based on the usage record of each word, and these methods may be followed. Thus, in step S54, the corresponding reading part word is obtained, and in step S55, the connected notation part word is obtained by tracing the pointer to the notation stored therein. The notation string stored therein is the notation corresponding to the desired reading, and in step S56 the notation string is output.

<Speech Synthesis> In speech synthesis, speech data corresponding to the notation data is retrieved from the notation data and speech synthesis is performed. FIG. 6 shows a flowchart of the processing procedure.
This flowchart is realized by the CPU 22 executing the program stored in the storage device 24 of FIG.

First, in step S61, a notation string is input from the input device 21. The notation string is entered as a character string containing kana and kanji. In step S62, the notation string is grammatically analyzed. In step S63, the word thus obtained is searched in the notation unit 34 of the dictionary 4 using the notation as a key. Thus, step S64
The notation word corresponding to the notation of the key is obtained at step S6
At 5, the pointer to the voice data stored therein is followed to obtain the concatenated voice part word. The voice data stored here is voice data corresponding to the desired notation, and the voice data is output as voice in step S66.

<Voice Recognition> In the voice recognition, the notation data corresponding to the voice data is searched for the voice recognition. FIG. 6 shows a flowchart of the processing procedure. This flowchart is realized by the CPU 22 executing the program stored in the storage device 24 of FIG.

First, in step S71, voice data is input from the input device 21. As the audio data, an audio signal converted into an electric signal by a microphone or the like is further digitized by a D / A converter or the like. Then, step S7
In step 2, necessary processing such as cutting out syllables is performed on the voice data. This processing may be similar to the conventionally known speech recognition processing. Regarding the syllable thus obtained, in step S73, step S72
The voice section 35 of the dictionary 4 is searched using the data obtained in step 1 as a key. Thus, in step S74, the voice part word corresponding to the voice data of the key is obtained, and in step S75, the pointer to the notation data stored therein is traced to obtain the connected notation part word. The notation string stored here is the notation string corresponding to the desired audio data, and step S76
Then, the notation data is output.

As described above, in the system of this embodiment,
It is possible to perform kana-kanji conversion / speech synthesis / speech recognition processing using one dictionary.

<Maintenance of dictionary> FIG. 24 is a flow chart showing an example of the procedure for the maintenance process of the dictionary 4.

First, the operator gives an instruction to start maintenance of the dictionary. The type of maintenance designated in step S241 is determined. There are three types of maintenance: addition, deletion, and modification.

If the word is to be deleted, the word to be deleted is entered in step S242. In order to input a word, it is sufficient to input the reading string. In step S243, the reading word can be obtained from the input reading, and the pointer can be followed to obtain the notation word and further the notation word from the notation word. . In the case of homophones, for example, all the notation strings of homophones may be displayed and the operator may be selected from them. The reading word, the notation word, and the voice word thus obtained are deleted from the dictionary.

In step S244, the index portion is updated to correspond to the deleted word.

On the other hand, in the case of change, step S245
Use to specify the word to be changed. This designation may be a reading designation. Next, in step S246, the portion to be changed is designated. The part to be changed is whether the reading is changed, the notation is changed, or the voice data is changed, and the contents after the change are changed.

If the changed contents are designated, in step S247, the target word, that is, the reading part word,
Read notation word, voice part word, necessary word,
The newly specified content is rewritten, and the read word is replaced with the rewritten word.

If it is an addition, step S2
Input the word to be added at 48. The additional words are reading,
Input notation and voice as a set. Then, the word added in step S249 is registered in the dictionary. In this case, it goes without saying that a pointer from the reading word to the notation word, a pointer from the notation word to the speech word, and a pointer from the speech word to the notation word are simultaneously written in each word.

Finally, the index is updated in step S250.

As described above, some attribute (reading in this embodiment) registered in the dictionary for the desired word is designated, the dictionary is searched from the designated attribute, and the registered content relating to the desired word is specified. By adding / deleting / changing all (reading part / notation part / voice data part in this embodiment), it is possible to maintain the consistency of the dictionary and reduce the maintenance time. According to the present embodiment, the reading, notation, and voice of words registered in the dictionary can be updated at a time. That is, maintenance of either the Kana-Kanji conversion dictionary or the voice synthesis / recognition dictionary is maintenance of the corresponding word in the other dictionary. Therefore, it is possible to reduce the trouble of maintaining the plurality of dictionaries and maintain the consistency between the plurality of dictionaries.

In the description of FIG. 24, in the case of deletion / change. Although the pronunciation is input, the notation or voice data can be input to maintain the corresponding word. In this case, since there is no pointer connected to the reading, the index will be used for the search.

As described above, the device of this embodiment is
Kana-to-Kanji conversion, voice synthesis, and voice recognition can be performed using one dictionary. In addition, since this dictionary is associated with pointers using reading, notation, and voice as headings, there is no overlap in the items in the dictionary, and minimal changes are required when maintaining the dictionary.

[0044]

Second Embodiment Next, an information processing system having a kana-kanji conversion / character recognition function, which is a second embodiment of the present invention, will be described.

The configuration of the information processing system of this embodiment is shown in the block diagram of FIG. 2 as in the first embodiment.

FIG. 8 is a block diagram of modules included in the system of this embodiment. As shown in FIG. 8, one dictionary 83 is used by the kana-kanji conversion unit 81 and the character recognition unit 82. The processing by each module is realized by the CPU 22 executing a program stored in the storage device 24. The dictionary 83 is stored in the storage device 24.

<Structure of Dictionary> FIG. 9 is a diagram showing the overall structure of an example of the dictionary 83. The dictionary 83 includes a character index 91, a character recognition data index 92, a reading unit 93, a notation unit 94, and a character recognition data unit 95. In the figure, a character index 91 is an index for retrieving the corresponding record from the key reading or notation, and a character recognition data index 92 is retrieving the corresponding record from the key character recognition data. It is an index for.

The reading part 93 is a part in which data for subtracting the notation from the kana or romaji reading is accumulated. As shown in FIG. 10A, the record (reading part word) has a reading string that serves as a heading, a pointer connected to a record (notation part word) indicating the corresponding notation, a part of speech, and a conjugation form. It stores grammatical information such as.

The notation section 94 is a section in which data for subtracting voice data corresponding to the notation such as Chinese characters is accumulated. The record (notation part word) is shown in FIG.
As shown in (b), a notation string as a heading,
The grammatical information of the record is stored.

The character recognition data part 95 is a part in which data for subtracting a notation corresponding to the data from the character recognition data is accumulated. As shown in FIG. 10C, the record (character recognition data section word) has a character recognition data to be a headline and a pointer to a record (notation section word) in which the corresponding notation is stored, The grammatical information of the record is stored. The character recognition data may be data representing features extracted from the character pattern.

<Kana-Kanji conversion> In Kana-Kanji conversion, the notation data corresponding to the reading data is searched from the reading data,
Kana-Kanji conversion is performed. FIG. 5 shows a flowchart of the processing procedure. Since this flowchart has been described in the first embodiment, the description will be omitted.

<Character Recognition> In character recognition, character recognition data is searched for corresponding notation data to perform character recognition. FIG. 11 shows a flowchart of the processing procedure.
This flowchart is realized by the CPU 22 executing the program stored in the storage device 24 of FIG.

First, in step S111, character image data is input from the input device 21 such as an image scanner. Next, in step S112, necessary processing such as extraction of characteristics of the character image data is performed to obtain character recognition data. This processing may be similar to the conventionally known character recognition processing. Regarding the character recognition data thus obtained, in step S113, the dictionary 83 is created using the data obtained in step S112 as a key.
The character recognition data section 95 is searched. Thus, in step S114, the character recognition data part word corresponding to the character recognition data of the key is obtained, and in step S115, the pointer to the notation data stored therein is traced to obtain the connected notation part word. The notation string stored here is the notation string corresponding to the desired character recognition data, and step S116
Then, the notation data is output.

With the above procedure, kana-kanji conversion / character recognition can be performed using one dictionary. The procedure for maintaining the dictionary is the same as that described with reference to FIG. 24 of the first embodiment. Therefore, the change of one dictionary is simultaneously changed of the other dictionaries, so that the consistency between the dictionaries can be maintained and maintenance work can be omitted.

[0055]

[Third Embodiment] Next, an information processing system having a kana-kanji conversion / kanji-kana conversion / kana-kanji mixed sentence analysis function according to a third embodiment of the present invention will be described.

The configuration of the information processing system of this embodiment is shown in the block diagram of FIG. 2 as in the first embodiment.

FIG. 12 is a block diagram of modules provided in the system of this embodiment. As shown in FIG. 12, one dictionary 124 is used by the kana-kanji conversion unit 121, the kanji-kana conversion unit 122, and the kanji-kana mixed sentence analysis unit 123. The processing by each module is realized by the CPU 22 executing a program stored in the storage device 24. The dictionary 124 is stored in the storage device 24.

<Dictionary Configuration> FIG. 12 is a diagram showing an example of the overall configuration of the dictionary 124. The dictionary 124 is composed of a character index part 131 and a word part 132. The word portion 132 is divided into a reading word portion and a notation word portion, each of which is composed of the reading word shown in FIG. 14A and the notation word shown in FIG. 14B. The reading word is for drawing a notation string with the reading string as a key, and the notation word is for drawing a reading with the notation string as a key. For example, the words are arranged by using the code in the reading column or the notation column as a key.

The contents of each word are shown in FIG. As shown in the figure, each word has grammatical information in addition to the correspondence between reading and notation. In this type of dictionary, the pointer cannot be traced for maintenance as in the first embodiment, but if the reading part word and the notation part word corresponding to each other are registered, either the reading or the notation can be performed. Both can be added / deleted / changed by specifying.

The structure of the dictionary may be as shown in FIG. In FIG. 15, the dictionary includes an index 151, a reading unit 152, and a writing unit 153. The reading part word 161 is shown in FIG. 16A, and has a reading string, a pointer to the corresponding notation part word, and grammar information. The notation word 162 has a notation string, a pointer to the corresponding reading word, and grammar information. The structure of this dictionary 124 is the same as the dictionary 4 of the first embodiment.
The voice part is removed from the configuration of the above, and the pointer from the notation part is not the voice part but the reading part. Therefore, the dictionary 124
The maintenance procedure is also performed in the same manner as the dictionary 4 of the first embodiment.

Furthermore, the structure of the dictionary 124 may be that shown in FIG. The dictionary of FIG. 17 is an example in which the reading column code and the notation code registered in the dictionary are different. For example, when a reading string or a word string is created according to the JIS code, one character requires 1 byte. However, if it is limited to the reading string, one byte for each character is sufficient if it is written in kana. Therefore,
The dictionary can be made smaller by representing the reading string with a 1-byte code and the notation string with a 2-byte code. The reading part word and the notation part word are respectively shown in FIG.
It is shown in (b). The configuration itself is the same as in Fig. 16, but
The reading string is described by the code A, and the notation string is described by another code B.

<Kana-Kanji conversion> In Kana-Kanji conversion, the notation data corresponding to the reading data is searched from the reading data,
Kana-Kanji conversion is performed. FIG. 5 shows a flowchart of the processing procedure. Since this flowchart has been described in the first embodiment, the description will be omitted. However, when the dictionary of FIG. 13 is used, it is not necessary to follow the pointer, and the display corresponding to reading is stored in the same record.

<Kanji-kana conversion> The kanji-kana conversion is for converting a sentence containing kanji into a reading string written in, for example, kana. FIG. 19 shows a flowchart of the processing procedure.

First, in step S191, the input device 21
Enter the notation sequence. The notation string is entered as a sentence containing mixed kanji. Next, in step S192, the notation string is analyzed such as cutting out words. In step S193, the dictionary 124 is searched for the words thus obtained using the code of the notation string as a key. Thus, the corresponding word is obtained in step S194, and step S195
Then, the reading string stored therein or the reading string connected by tracing the pointer is obtained. This reading string is a reading corresponding to the notation required, and the reading string is output in step S196.

<Kanji / Kana Sentence Analysis> The Kanji / Kana mixed sentence analysis obtains grammatical information corresponding to the written data or reading data and analyzes the sentence. FIG. 20 shows a flowchart of the processing procedure.

First, in step S201, the input device 21
Enter the notation sequence. The notation string is entered as a sentence containing mixed kanji. Next, in step S202, the dictionary is searched by the notation string. If the word of Kana is the target, the dictionary may be searched from the reading string. Thus, step S203
The word corresponding to the word string or the reading string to be searched can be acquired with, and the grammatical information stored in the word acquired in step S203 can be obtained. Using this grammar information, the notation string input in step S205 is analyzed.

With the above procedure, kana-kanji conversion, kanji-kana conversion, and kanji-kana sentence analysis can be performed using one dictionary. The procedure for maintaining the dictionary is the same as that described with reference to FIG. 24 of the first embodiment. Therefore, changes in one dictionary will be changes in the other dictionary at the same time,
Uniformity between dictionaries can be maintained and maintenance work can be omitted.

[0068]

Fourth Embodiment Next, an information processing system having a kana-kanji conversion / voice synthesis / voice recognition / character recognition function, which is a fourth embodiment of the present invention, will be described.

The configuration of the information processing system of this embodiment is shown in the block diagram of FIG. 2 as in the first embodiment.

FIG. 21 is a block diagram of modules included in the system of this embodiment. As shown in FIG. 21, one dictionary 215 is used as a kana-kanji conversion unit 211 and a voice synthesis unit 2.
12, used by the voice recognition unit 213 and the character recognition unit 214. The processing by each module is realized by the CPU 22 executing a program stored in the storage device 24. The dictionary 215 is also stored in the storage device 24.

<Structure of Dictionary> FIG. 22 shows an example of the overall structure of the dictionary 215. The dictionary 215 includes a character index unit 225, a voice index unit 226, a reading unit 221, and a writing unit 2.
22 and a voice unit 223 and a character recognition data unit 224. The reading unit 221 stores a record (word) whose heading is reading written in kana or the like,
The reading and the notation are associated with each other. The notation section 222 stores records having headings for notations such as Chinese characters.
Corresponds to notation and voice. The voice section 223 stores a record having voice data as a headline, which is associated with the notation. The character recognition data unit 224 stores character recognition data, which is associated with the notation.

The contents of each word are shown in FIG.
It is shown in (d). Each word stores a headline, a pointer to the corresponding word, and grammar information. Kana-Kanji conversion / speech synthesis using this dictionary 215
The voice recognition / character recognition processing is the same procedure as described in the first and second embodiments, and the description thereof will be omitted, but using one dictionary, kana-kanji conversion / voice synthesis / voice recognition / character Can recognize. The procedure for maintaining the dictionary is the same as that described with reference to FIG. 24 of the first embodiment. For this reason, the change of one dictionary becomes the change of the other dictionary at the same time, and the consistency between the dictionaries can be maintained, and the trouble of maintenance can be saved.

As described above, various dictionaries are combined into one and used in the same manner as in the first to fourth embodiments.
It can be maintained, thereby maintaining the integrity of the dictionary, simplifying the maintenance of the dictionary and reducing the size of the dictionary. Further, not only words but the same items are registered. The present invention may be applied to a system including a plurality of devices or an apparatus including a single device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0074]

As described above, the document processing apparatus according to the present invention can reduce the capacity of the dictionary used for a plurality of purposes, maintain the uniformity of the contents, and facilitate the maintenance. effective.

[0075]

[Brief description of drawings]

FIG. 1 is a block diagram illustrating a configuration of a document processing system according to a first embodiment.

FIG. 2 is a block diagram showing a configuration of an information processing system that is an embodiment of the present invention.

FIG. 3 is an overall configuration diagram of a dictionary according to a first embodiment.

FIG. 4 is a diagram showing a structure of words in a dictionary according to the first embodiment.

FIG. 5 is a flowchart of a kana-kanji conversion processing procedure.

FIG. 6 is a flowchart of a voice synthesis processing procedure.

FIG. 7 is a flowchart of a voice recognition processing procedure.

FIG. 8 is a block diagram showing a configuration of a document processing system according to a second exemplary embodiment.

FIG. 9 is an overall configuration diagram of a dictionary according to a second embodiment.

FIG. 10 is a diagram showing a structure of words in a dictionary according to a second embodiment.

FIG. 11 is a flowchart of a character recognition processing procedure.

FIG. 12 is a block diagram showing a configuration of a document processing system of a third exemplary embodiment.

FIG. 13 is an overall configuration diagram of a dictionary according to a third embodiment.

FIG. 14 is a diagram showing a structure of words in a dictionary according to a third embodiment.

FIG. 15 is an overall configuration diagram of a dictionary having another configuration according to the third embodiment.

16 is a diagram showing a structure of words in the dictionary shown in FIG. 15. FIG.

FIG. 17 is an overall configuration diagram of a dictionary having another configuration according to the third embodiment.

FIG. 18 is a diagram showing a structure of words in the dictionary shown in FIG. 17;

FIG. 19 is a flowchart of a kanji-kana conversion processing procedure.

FIG. 20 is a flowchart of a kanji / kana sentence analysis processing procedure.

FIG. 21 is a block diagram showing the configuration of a document processing system according to a fourth embodiment.

FIG. 22 is an overall configuration diagram of a dictionary according to a fourth embodiment.

FIG. 23 is a diagram showing a structure of words in a dictionary according to a fourth embodiment.

FIG. 24 is a flowchart showing a processing procedure of dictionary maintenance. 1 Kana-Kanji conversion unit, 2 voice synthesis unit, 3 voice recognition unit, 4 dictionary, 21 input device, 22 central processing unit, 23 output device, 24 storage device.

Claims (6)

[Claims] 1. A dictionary in which a plurality of attributes related to a word are registered as headings of corresponding words, and a plurality of attributes related to the same word are associated with each other, and other words corresponding to predetermined attributes of a desired word A document processing apparatus comprising: a search unit that searches the attribute of the item from the dictionary; and a unit that performs a subsequent process based on the attribute searched by the search unit. 2. A designating unit for designating maintenance for the dictionary, and a maintenance unit for maintaining the attribute of the word retrieved by the retrieval unit as designated by the designating unit.
The document processing apparatus according to claim 1, further comprising: 3. The designating means designates either addition or deletion of a word or change of attribute, and designation of an attribute of a target word, and the maintenance means is designated in case of addition. Registered attributes as headings in the dictionary, in the case of deletion, all attributes related to the word corresponding to the specified attribute are deleted from the dictionary, and in case of change, the specified attribute is changed to the specified value. The document processing apparatus according to claim 2, wherein: 4. The plurality of attributes include reading, notation, and voice data, and the processing in the next stage includes kana-kanji conversion, voice synthesis, and voice recognition. The document processing device according to any one of the above. 5. The method according to claim 1, wherein the plurality of attributes include reading, notation, and character recognition data, and the processing in the next stage includes character recognition and kana-kanji conversion. Document processing device described. 6. The plurality of attributes include reading and notation, voice data, and character recognition data, and the processing in the next stage includes kana-kanji conversion, voice synthesis, voice recognition, and character recognition. The document processing apparatus according to claim 1.